1)To overcome hetero or auto correlation, we use xtgls, xtregar ( Cross-sectional time-series FGLS regression ) in panel data? Should I run any other test or need?
2)Which test should I run for unbalanced panel data ? For example Fixed and random effect,GMM ,Probit and logit regression or Pooled OLS etc.
(N=89,T=15)
Thanks in advance